Vehicle Collision Avoidance Using Multi-Point Threat Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicle collision mitigation at intersections is challenging due to limited or inaccurate data leading to excessive risk assessment and operational costs, with existing systems often resulting in false positive identifications and unnecessary processing cycles.
Innovation Solution
A system that determines a rear time to collision and overall threat number based on both front and rear threat numbers, using a threat multiplier to adjust the risk assessment and actuate components such as brakes or steering to avoid collisions, while preventing unnecessary braking during near miss scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If threat assessment is performed using limited or inaccurate data, then risk assessment speed is improved, but measurement precision deteriorates leading to false positives
Solution Approach 1:
The patent segments the target vehicle into multiple points (front point, rear point, and intermediate points) along its longitudinal axis. Each point is independently assessed for collision risk, allowing the system to evaluate different portions of the target vehicle separately. This segmentation enables more precise threat assessment by identifying which specific portions pose actual collision risks, rather than treating the entire target vehicle as a single threat unit.
Solution Approach 2:
The patent introduces a longitudinal dimension to threat assessment by evaluating multiple points along the length of the target vehicle. Instead of a single two-dimensional collision risk assessment, the system adds the longitudinal dimension by assessing front, rear, and intermediate points, creating a more comprehensive three-dimensional threat model that reduces false positives while maintaining processing efficiency.
2Reliability
If threat assessment assigns excessive risk to scenarios, then collision avoidance is improved, but operational cost increases due to unnecessary braking
Solution Approach 1:
By segmenting the target vehicle into multiple assessment points, the system can identify that only certain portions (e.g., front point) pose collision risks while other portions (e.g., rear point) do not. This allows the system to avoid unnecessary braking actions that would be triggered by treating the entire target vehicle as a uniform threat, thereby reducing energy loss from false positive braking events.
Solution Approach 2:
The patent applies local quality assessment by evaluating different points of the target vehicle independently with different threat levels. The front point may be assessed as high risk while the rear point is assessed as low or no risk. This localized assessment allows the system to trigger avoidance actions only when locally relevant threats are detected, preventing unnecessary global braking responses and reducing energy consumption.
3Measurement precision
If multiple points of target information are used, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent divides the target vehicle into a manageable number of discrete points (front point, rear point, and selected intermediate points) rather than continuously analyzing the entire vehicle surface. This segmentation approach maintains measurement precision by capturing critical threat information at key locations while limiting processing complexity through a finite, structured set of assessment points.
Solution Approach 2:
The system performs partial assessment by focusing on specific critical points of the target vehicle rather than comprehensively analyzing every possible location. This partial action approach achieves sufficient measurement precision for collision avoidance by concentrating computational resources on the most relevant points, thereby managing device complexity while maintaining adequate threat assessment accuracy.
Data Source
AI summary
A system includes a computer including a processor and a memory, the memory storing instructions executable by the processor to, upon determining a front threat number exceeds a threat threshold, determine a rear time to collision between a turning host vehicle and a target, and, upon determining that the rear time to collision is below a time threshold, actuate a component based on a rear threat number for the target.